19
TimewARP • User Guide
English
3.4.1.1
How Adding Pitches Means Multiplying Frequencies
The range of audio frequencies—of human hearing, in other words—is conventionally stated to be 20Hz to 20KHz.
And when you think linearly, it sounds tragic to learn of someone, say, who can only hear up to 10kHz. But in the
realm of human pitch perception, such a person has kept 90% of his hearing range: from 10kHz to 20kHz is just
one octave out of the 10 we hear.
This is a really important fact in audio synthesis; you will encounter it over and over again, in every patch you
create. It governs much of the arithmetic of generating and controlling spectral distributions and patches:
• Equal pitch intervals require exponentially increasing frequency increments;
• Equal frequency increments make a harmonic series. In a harmonic series, as we go up the series we get
smaller and smaller pitch intervals.
3.4.2
Signal Amplitude and Audible Volume
This is a very dicey relationship. It is true that, for any given signal, increasing its amplitude will increase the
volume of the associated sound. But a lot of other signal characteristics have a greater impact on our perception
of volume than amplitude does.
For example, the difference between talking and shouting at someone is far more a matter of pitch and spectrum—
tone-color—than of mere amplitude. When I yell, I “raise my voice”; that is, I raise the pitch of my voice, and I put
more energy into it, which generates more harmonics, which is a matter of spectral content, not amplitude. Stage
actors have to learn to overcome this tendency; in order to be heard onstage without shouting, they have to learn
to increase their speaking amplitude without yelling.
If you examine—visually—the recorded signal from a pipe organ, or even an orchestra, you may be surprised
to find the apparent amplitude of the softer signals almost equal to that of the loudest. This has to do with the
spectral distribution of the sound. Of two different signals of approximately equal amplitude, the one with the
broadest spectral distribution—the most harmonic content—will sound louder.